We propose to investigate the effect of selenium intake on neoplastic polyp formation, an intermediate stage of human colorectal carcinogenesis. We propose to conduct a nested case-control study using data and stored serum samples from a large, multi-centered polyps prevention trial designed to test the efficacy of oral betacarotene, vitamin C and vitamin E supplementation in reducing adenoma occurrences. The study was comprised of patients with a history of adenoma from six clinical centers. At study entry, participants completed a health habits and history questionnaire and provided dietary intake data (using a standardize food frequency questionnaire). Medical history information was updated semi- annually. Serum samples were collected under mineral-free conditions throughout the study. All participants were free of polyps at study entry and underwent a colonoscopic evaluation after one year to ensure that no polyps were missed,and after four years to identify new adenoma occurrences. All biopsies were reviewed by two pathologists for the presence of neoplasia. For the proposed nested case-control study, we will select as cases, all 278 participants who had an adenoma occurrence between the year one and year four examinations. For comparison, we will randomly choose and equal number of participants who completed the year four examinations but had not developed any adenoma. Prediagnostic serum selenium concentrations (an indicator of dietary selenium) will be determined using instrumental neutron activation analysis. We will estimated the relative risk of adenoma occurrence associated with serum selenium concentrations using stratified and logistic regression analysis. We will further assess the relation between serum selenium levels and the number, size and location of adenoma occurrences. A major advantage of our study design is that patients were followed with endoscopic examination at prescribed intervals, rather than for symptoms or based on patient or physician preferences. Therefore, our results will not be subject to the detection bias which cannot be avoided in most longitudinal studies. Also, the proposed study can be conducted in considerable less time and expense than initiating a new study of this size. Thus, our findings should contribute valuable insights for planning future intervention studies with chemopreventive and nutritional agents, and to the understanding of large bowel carcinogenesis and its prevention.